Search papers, labs, and topics across Lattice.
5
0
7
10
LLMs can robustly shift simulated stances in online discussions, but their sensitivity to context changes raises critical questions about reliability in opinion dynamics.
LLMs can exploit societal regulations, discovering loopholes that allow them to circumvent intended compliance while appearing to follow the rules.
Personalizing LLMs through a sociologically grounded framework reveals the hierarchical nature of user behavior, leading to significant performance gains across tasks.
Escape the communication trilemma: HyLaT achieves efficient, interpretable multi-agent communication by strategically blending latent-space and natural language channels.
Current LLMs fall short in understanding implicit intentions and modeling long-term user preferences, as revealed by a new benchmark, LifeSim-Eval, designed to simulate real-world user-assistant interactions.
